DS-InSAR 计算result.zip

采矿沉降是煤盆地普遍存在的地质灾害，精确、可靠的变形监测对于有效降低风险至关重要。传统的时间序列 InSAR 存在植被诱导的去相关、裸露地表散射和大气延迟，从而降低了相干像素;此外，大多数预测模型仅利用时间信息。为了解决这些限制，我们引入了一个集成的 DS InSAR + CNN LSTM 框架，用于沉降监测和预报。使用 GACOS 数据校正的 43 个 Sentinel 1A 场景 （2017-2018） 进行了处理，得出累积变形，根据多视图 SBAS InSAR 进行交叉验证，并用于训练提前一年预测趋势的 CNN LSTM 网络。

Mining subsidence is a prevalent geological hazard in coal basins, and accurate and reliable deformation monitoring is crucial for effectively reducing risks. Traditional time-series InSAR suffers from vegetation-induced decorrelation, bare surface scattering, and atmospheric delay, which reduce coherent pixels; furthermore, most prediction models only utilize temporal information. To address these limitations, we introduce an integrated DS InSAR + CNN LSTM framework for subsidence monitoring and prediction. Forty-three Sentinel-1A scenes (2017-2018) corrected with GACOS data were processed to yield cumulative deformation, which was cross-validated against multi-view SBAS InSAR and used to train a CNN LSTM network that predicts trends one year in advance.